Description
Presents results which form a continuation of studies on a neural-network-based controller. Summarises the previous studies. Presents new results and offers some novel techniques for training the network and making the entire package easier to use. The two major additions are – 1) efficient use of training data for dramatically reducing memory requirements and 2) incorporation of a PID algorithm for performing control during training periods.
KEYWORDS: year 1996, Controls, expert systems, algorithms
Citation: Symposium Papers, Atlanta, GA, 1996
Product Details
- Published:
- 1996
- Number of Pages:
- 12
- File Size:
- 1 file , 1 MB
- Product Code(s):
- D-17206
